Gromiha, M. M. (2005). A statistical model for predicting protein folding rates from amino acid sequence with structural class information. Journal of chemical information and modeling, 45(2), 494–501. http://dx.doi.org/10.1021/ci049757q
Introduction
-Issue at hand is predicting 3-D structure of protein from AA sequence.
-Found long-range order is one of the best parameters to handle this problem
-3 groups have worked on this issue each proposing different models with using multiple states
-This paper's goal is develop a model for predicting protein folding rates from amino acid sequences
Materials and Methods
-Experimental ln(kf) values were utilized for experimental folding rates
-Amino acid properties used were physical-chemical, energetic and conformational
-Average amino acid property for each protein computed utilizing the standard formula
Results and Discussion
-Found classifying alpha and beta proteins enhanced the correlations from 0.39 to 0.97
-Able to descriminate the properties from random values from good correlations with all alpha proteins
-Developed a regression model for predicting folding rates
-Performed a jack-knife test and determined coefficients of regression equation
-Similar calculations for Beta proteins provided a less accurate method for prediction
-Correlation coefficients for a different combination of 4 AA properties were analyzed
-Predicting folding rates had a good agreement with experimental data. Correlated 0.956
-Analyzed coefficients for mixed AA proteins and found weaker correlation for predictions of properties
-Utilized the 3 derived equations for formulate one for protein folding rates
-Found correlation of 0.97 with predicted and experimental
-Repeated for 2 and 3 state proteins and achieved correlations above 0.90
-Worked on folding rates for proteins with mutations
-Slightly lower correlations were achieved. Around 0.87
-Compared correlations to other groups calculations and formulas
-CO - 0.74
-LRO - 0.81
-TCD - 0.88
-Aknowledged direct comparison of coefficients in present work with others work is not appropriate
Conclusions
-Amino acid residues dictate protein structure based on their interactions with surrounding medium
-Linear regression models for predicting folding rates of 2 and 3 state proteins were formulated
-Folding rates showed excellent correlations with experimental
Introduction
-Issue at hand is predicting 3-D structure of protein from AA sequence.
-Found long-range order is one of the best parameters to handle this problem
-3 groups have worked on this issue each proposing different models with using multiple states
-This paper's goal is develop a model for predicting protein folding rates from amino acid sequences
Materials and Methods
-Experimental ln(kf) values were utilized for experimental folding rates
-Amino acid properties used were physical-chemical, energetic and conformational
-Average amino acid property for each protein computed utilizing the standard formula
Results and Discussion
-Found classifying alpha and beta proteins enhanced the correlations from 0.39 to 0.97
-Able to descriminate the properties from random values from good correlations with all alpha proteins
-Developed a regression model for predicting folding rates
-Performed a jack-knife test and determined coefficients of regression equation
-Similar calculations for Beta proteins provided a less accurate method for prediction
-Correlation coefficients for a different combination of 4 AA properties were analyzed
-Predicting folding rates had a good agreement with experimental data. Correlated 0.956
-Analyzed coefficients for mixed AA proteins and found weaker correlation for predictions of properties
-Utilized the 3 derived equations for formulate one for protein folding rates
-Found correlation of 0.97 with predicted and experimental
-Repeated for 2 and 3 state proteins and achieved correlations above 0.90
-Worked on folding rates for proteins with mutations
-Slightly lower correlations were achieved. Around 0.87
-Compared correlations to other groups calculations and formulas
-CO - 0.74
-LRO - 0.81
-TCD - 0.88
-Aknowledged direct comparison of coefficients in present work with others work is not appropriate
Conclusions
-Amino acid residues dictate protein structure based on their interactions with surrounding medium
-Linear regression models for predicting folding rates of 2 and 3 state proteins were formulated
-Folding rates showed excellent correlations with experimental